Automating Surgical Quality Data Shows High Accuracy

Using electronic health record (EHR) data for surgical quality improvement matches manual abstraction accuracy, easing the burden on resources.

• Strong agreement for race and ethnicity variables (kappa = 98.1%).
• Preoperative risk factors show lower agreement (median kappa = 28.6%), indicating caution in reliance on EHR for this domain.

This shift could streamline data collection, enhancing surgical outcomes monitoring.

• Postoperative complications had a median kappa of only 15.1%, highlighting areas needing manual review.